Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Can we have a provision to reuse the reflatent? #13

Open
kris775 opened this issue Feb 16, 2024 · 3 comments
Open

Can we have a provision to reuse the reflatent? #13

kris775 opened this issue Feb 16, 2024 · 3 comments

Comments

@kris775
Copy link

kris775 commented Feb 16, 2024

Can we save the reference latent generated by the node "Style aligned sample reference latent' for one style and then reuse it again in "Style aligned reference sampler"?.
If the processing of reference image need not be done again for each run when using the same reference image then saving and reusing might help in reducing the time of each run a bit and can help in a continuous iteration type of workflow.

@jasonphillips
Copy link
Contributor

Do you mean saving it offline / on disk for future use, or do you just mean caching the result between runs?

It should already cache and not require a re-run between generations if you use a fixed seed for it and the same image input / prompt input to the node, so perhaps I'm misunderstanding the request.

@dfl
Copy link

dfl commented Feb 21, 2024

that idea reminds me a bit of this feature... saving embeds for IP adapter https://youtu.be/mJQ62ly7jrg?si=V37H7HMK_KeymnPl&t=846

@kris775
Copy link
Author

kris775 commented Feb 22, 2024

Do you mean saving it offline / on disk for future use, or do you just mean caching the result between runs?

Yes i mean saving it offline on disk and reusing later. its like having pre saved different styles which can be directly used, similar to saving embeds for the IP adapter.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants